JP6927308B2 - Voice control device and its control method - Google Patents

Description

The present invention relates to a voice operating device having a speaker identification function.

In recent years, AI speakers (also called smart speakers) having a voice assistant function capable of being operated by voice have been attracting attention. According to such a voice operation device, desired information retrieval and operation of linked home appliances can be performed by voice, so that the user can perform tasks such as housework, cleaning, and washing without using his / her hands. You can perform operations. In addition, since the voice control device can be realized at low cost with a simple configuration that uses a microphone, a speaker, and a processor for voice recognition as the minimum components, it can be easily arranged in a plurality of rooms. ..

The voice operation device described in Patent Document 1 has a speaker identification function for identifying who the speaker is from the voice quality of the voice, in addition to the voice recognition function for recognizing the voice operation by the user. This makes it possible to provide services that match the tastes of users.

Japanese Unexamined Patent Publication No. 2010-286702

However, although the speaker identification technology is being researched and developed, the identification accuracy is still not high. Especially for AI speakers, it is desirable to make the operation voice atypical in order to avoid spoofing using the recorded operation voice, but if the operation voice is atypical and short, identify the speaker. Is especially difficult. In addition, since the voice quality of the voice emitted by the user changes greatly depending on the distance of the sound source, the ambient temperature, the environment such as reflection, etc., the operation voice is compared with a single voice quality model of the user registered in advance. It is difficult to identify the speaker by itself. Therefore, an object of the present invention is to provide a voice operation device capable of further improving the accuracy of speaker identification and a control method thereof.

According to one aspect of the present invention, a speaker identification unit that identifies a user as a speaker for voice operation based on voice information and a user's voice quality model registered in advance, and a voice recognition unit that recognizes voice information for voice operation. It is a voice operation device including a voice operation recognition unit that generates information, and the speaker identification unit includes voice operation information, position information of the voice operation device, direction information of the speaker, distance information of the speaker, and time. A voice control device that identifies a speaker by using at least one of the information as auxiliary information is provided.

Further, according to another viewpoint of the present invention, a speaker identification unit that identifies the user as a speaker of voice operation based on the voice information and a voice quality model of the user registered in advance, and voice recognition of the voice information are performed. It is a control method of a voice operation device including a voice operation recognition unit for generating voice operation information, and the speaker identification unit is a voice operation information, a position information of the voice operation device, a speaker's direction information, and a speaker. Provided is a control method of a voice operation device having a step of identifying a speaker by using at least one of the distance information and the time information of the above as auxiliary information.

According to the present invention, it is possible to provide a voice operation device capable of further improving the accuracy of speaker identification and a control method thereof.

It is a block diagram which shows schematic structure of the voice operation apparatus which concerns on 1st Embodiment. It is a figure which shows the arrangement example of the voice operation apparatus which concerns on this invention. It is a flowchart which shows the control method of the voice operation apparatus which concerns on 1st Embodiment. It is a block diagram which shows schematic structure of the voice operation apparatus which concerns on 2nd Embodiment. It is a flowchart which shows the control method of the voice operation apparatus which concerns on 2nd Embodiment. It is a block diagram which shows schematic structure of the voice operation apparatus which concerns on 3rd Embodiment. It is a block diagram which shows schematic structure of the voice operation apparatus which concerns on 4th Embodiment. It is a block diagram which shows schematic structure of the voice operation apparatus which concerns on 5th Embodiment. It is a block diagram which shows schematic structure of the voice operation apparatus which concerns on 6th Embodiment.

Hereinafter, preferred embodiments of the present invention will be described with reference to the drawings. The present invention is not limited to the following embodiments, and can be appropriately modified without departing from the gist thereof. Those having the same or corresponding functions in each figure are designated by the same reference numerals, and the description thereof may be omitted or simplified.

(First Embodiment)
FIG. 1 is a block diagram schematically showing a configuration of a voice operating device according to the first embodiment. FIG. 1 shows an example in which the voice control device of the present embodiment is applied to an AI speaker. This figure is only an example, and the voice operation device of the present embodiment can be applied to, for example, a robot, a smartphone, a car navigation system, or the like that can recognize a voice operation by a user. The voice operation device of the present embodiment includes a speaker 1, a microphone 2, a voice output unit 11, a voice input unit 21, and a control calculation unit 3. In a typical AI speaker, these components are housed in the housing of the speaker 1.

The control calculation unit 3 is a processor such as a CPU that executes a program stored in a storage medium and controls and calculates each component of the voice operation device of the present embodiment. The control calculation unit 3 includes a storage medium (not shown) in which the program is stored. The control calculation unit 3 includes a voice operation response unit 31, a voice operation recognition unit 32, a speaker identification unit 33, a voice quality model 331, and a wireless communication unit 35.

The microphone 2 converts the voice vibration caused by the user's voice operation into an electric signal. The voice input unit 21 converts the analog signal input from the microphone 2 into digital voice information, and performs voice signal processing on the voice information. Here, the voice input unit 21 may be configured to be included in the microphone 2.

The voice operation recognition unit 32 voice-recognizes the voice information input from the voice input unit 21 and generates voice operation information. The control calculation unit 3 executes an operation process corresponding to the generated voice operation information. Specifically, desired information retrieval and operation processing of linked home appliances are performed via the wireless communication unit 35.

After that, the voice operation response unit 31 notifies the user of the execution result of the operation process via the voice output unit 11 and the speaker 1. Specifically, the voice output unit 11 converts the digital information input from the voice operation response unit 31 into an analog signal. The speaker 1 converts the electric signal input from the voice output unit 11 into voice vibration and outputs it. Here, the audio output unit 11 may be configured to be included in the speaker 1.

The registered user information 330 is stored in the storage medium of the control calculation unit 3. The voice quality model 331 of the user is registered in advance in the registered user information 330. The speaker identification unit 33 calculates the degree of similarity between the voice information input from the voice input unit 21 and the user's voice quality model 331, and identifies the user with the highest degree of similarity as the speaker of the voice operation. As a result, it is possible to provide a service according to the user's preference, and it is possible to further improve the convenience of the user. For example, information according to the user's preference is registered in the registered user information 330 in advance, and operation processing corresponding to the voice operation information is executed according to the user's preference, or information according to the user's preference can be executed in advance in the registered user information 330. It is possible to prohibit voice operations by anyone other than the registered user.

However, as mentioned above, the speaker identification technology is still not highly accurate, although research and development are progressing. Therefore, the speaker identification unit 33 of the present invention uses at least one of voice operation information, position information of the voice control device, direction information of the speaker, distance information of the speaker, and time information as auxiliary information. It is characterized by identifying a person. For example, in the voice operation device of the present embodiment shown in FIG. 1, the voice operation recognition unit 32 is used as the auxiliary information calculation unit 34 for calculating the auxiliary information.

FIG. 2 is a diagram showing an arrangement example of the voice operation device according to the present invention. The voice control device according to the present invention is arranged at the position of the microphone 2 shown in FIG. It is assumed that the users 61 and 62 existing in the room of FIG. 2 are, for example, a mother and a father relaxing in the living room at home.

In this case, the speaker identification unit 33 of the present embodiment determines that there is a high possibility that the speaker of the voice operation is the mother, for example, when the instruction by the voice operation is the acquisition of the weather information. Alternatively, when the instruction by voice operation is the acquisition of stock price information, it is determined that the speaker of voice operation is likely to be a father. Then, the speaker identification unit 33 corrects the similarity with the user's voice quality model 331 according to the voice operation information.

Here, information such as keywords characteristic of the user may be registered in the registered user information 330 in advance in association with the keywords included in the voice operation information, or the frequency of the keywords used by the user may be changed. You may study at any time based on this. When the voice operation device learns, the correction value of the similarity is registered in advance in the registered user information 330 for each user, and the result of identifying the speaker is reflected in the correction value. As a result, the speaker identification unit 33 learns the correlation between the auxiliary information characteristic of the user and the correction value of the similarity. Therefore, each time the speaker performs an operation process corresponding to the voice operation, the speaker The accuracy of identification can be improved.

FIG. 3 is a flowchart showing a control method of the voice operation device according to the first embodiment. In step S101, the microphone 2 and the voice input unit 21 convert the voice vibration caused by the user's voice operation into an electric signal. In step S102, the voice operation recognition unit 32 voice-recognizes the voice information input from the microphone 2 and the voice input unit 21 and generates the voice operation information. Here, a well-known technique can be used as the voice recognition method performed by the voice operation recognition unit 32.

In step S103, the speaker identification unit 33 reads out the voice quality model 331 of the user registered in advance in the registered user information 330 together with the registered user information 330. In step S104, the speaker identification unit 33 calculates the degree of similarity between the voice information input from the voice input unit 21 and the user's voice quality model 331.

Here, a well-known technique can be used as the speaker recognition method performed by the speaker identification unit 33. _{For example, the distance s between the voice waveform f 0} of the user's voice quality model 331 registered in advance in the registered user information 330 and _{the voice waveform f 1} of the voice information input from the voice input unit 21 is calculated by the following equation (1). Find it and use its inverse 1 / s as the similarity. Here, Σ is assumed to be summed at a plurality of time t within a predetermined period. Further, it is assumed that the voice waveforms f ₀ and f _{1 are appropriately normalized.}
s ² = Σ {f ₁ (t) -f ₀ (t)} ² (1)

_{Alternatively, the distance s between the frequency spectrum F 0} of the user's voice quality model 331 registered in advance in the registered user information 330 and _{the frequency spectrum F 1} of the voice information input from the voice input unit 21 is calculated by the following equation (2). The reciprocal 1 / s may be used as the degree of similarity. Here, Σ is assumed to be summed at a plurality of frequencies k within a predetermined range. Further, it is assumed that the frequency spectra F ₀ and F _{1 are appropriately normalized.}
s ² = Σ {F ₁ (k) -F ₀ (k)} ² (2)

In step S105, the speaker identification unit 33 corrects the similarity obtained by the above equation (1) or (2) according to the keyword included in the voice operation information. For example, when the instruction by voice operation is the acquisition of weather information, the similarity 1 / s of the voice waveform of the mother's voice quality model 331 is increased by +5 points. Alternatively, when the instruction by voice operation is the acquisition of stock price information, the similarity 1 / s of the voice waveform of the father's voice quality model 331 is increased by +5 points. Here, the magnitude of the correction value of the similarity can be appropriately determined for each keyword according to the application, the user, and the like.

In step S106, the speaker identification unit 33 identifies the user with the highest degree of similarity after correction as the speaker of the voice operation. In step S107, the control calculation unit 3 executes the operation process corresponding to the voice operation information according to the preference of the identified user. In step S108, the voice operation response unit 31 notifies the user by voice of the result of the process executed in step S107 via the speaker 1 and the voice output unit 11.

As described above, in the voice operation device of the present embodiment, the registered user information includes keywords characteristic to the user, and the speaker identification unit corrects the similarity according to the keywords included in the voice operation information. .. According to such a configuration, it is possible to provide a voice operation device capable of further improving the accuracy of speaker identification and a control method thereof.

(Second Embodiment)
FIG. 4 is a block diagram schematically showing the configuration of the voice operation device according to the second embodiment. The voice operation device of this embodiment is different from the previous first embodiment in that it includes a GPS device 41 and a position calculation unit 341. Other configurations are substantially the same as those of the first embodiment. Hereinafter, a configuration different from that of the first embodiment will be mainly described.

The GPS device 41 can acquire GPS (Global Positioning System) information. The position calculation unit 341 calculates the position information in which the voice operation device is arranged by using the GPS information input from the GPS device 41. Here, the position calculation unit 341 may be configured to be included in the GPS device 41. As described above, the voice operation device of the present embodiment is characterized by including the position calculation unit 341 as the auxiliary information calculation unit 34 for calculating the auxiliary information.

As described above, the voice quality of the voice emitted by the user changes greatly depending on the environment such as the distance of the sound source, the ambient temperature, and the reflection. Therefore, the operation voice is a single voice quality model of the user registered in advance. It is difficult to identify the speaker just by comparing with 331.

Therefore, the registered user information 330 of the present embodiment has a plurality of voice quality models 331 as a database for each user. Then, the speaker identification unit 33 selects the voice quality model 331 according to the position information of the voice operation device calculated by the position calculation unit 341. For example, when the voice control device is arranged at home, the voice quality model 331 suitable for the home environment is selected. Alternatively, if the voice control device is located in the workplace, the voice quality model 331 suitable for the workplace environment is selected. Alternatively, when the voice control device is arranged in the moving vehicle, the voice quality model 331 suitable for the environment in the vehicle is selected. These plurality of voice quality models 331 are registered in advance in the registered user information 330 in association with the position information of the voice operation device.

As a result, even when the voice operation device of the present embodiment is applied to a movable device such as a robot, a smartphone, a car navigation system, etc., the voice quality model 331 suitable for the surrounding environment is selected, so that the speaker The accuracy of identification can be further improved.

FIG. 5 is a flowchart showing a control method of the voice operation device according to the second embodiment. In the flowchart of the present embodiment shown in FIG. 5, when the voice quality model 331 is read out in step S203, the voice quality model 331 is selected according to the position information of the voice operating device. It is different from 3. In the flowchart of FIG. 5, the similarity correction process shown in step S105 of FIG. 3 of the first embodiment is omitted, but in this embodiment as well, step S105 may be performed in the same manner as in FIG. ..

As described above, the voice control device of the present embodiment further includes a GPS device that acquires the position of the voice control device as position information. Then, the registered user information has a plurality of voice quality models for each user, and the speaker identification unit selects the voice quality model according to the position information. Even with such a configuration, it is possible to provide a voice operation device and a control method thereof that can further improve the accuracy of speaker identification.

(Third Embodiment)
FIG. 6 is a block diagram schematically showing the configuration of the voice operation device according to the third embodiment. The voice operation device of the present embodiment is different from the first embodiment in that it includes an array microphone 42 and a voice direction calculation unit 342. Other configurations are substantially the same as those in the first embodiment. Hereinafter, a configuration different from that of the first embodiment will be mainly described.

The array microphone 42 has a configuration in which a plurality of microphones are arranged, and it is possible to specify the direction of a speaker for voice operation. The voice direction calculation unit 342 calculates the direction information of the speaker as seen from the voice control device based on the information input from the array microphone 42. Here, the voice direction calculation unit 342 may be included in the array microphone 42. As described above, the voice operation device of the present embodiment is characterized by including the voice direction calculation unit 342 as the auxiliary information calculation unit 34 for calculating the auxiliary information.

For example, in the room shown in FIG. 2, the chair on the left side when viewed from the voice control device is the seat for the mother, and the chair on the right side is the seat for the father. In this case, the speaker identification unit 33 of the present embodiment determines that, for example, when the direction of the voice-operated speaker is on the left side, the voice-operated speaker is likely to be the mother. Alternatively, when the direction of the voice-operated speaker is on the right side, it is determined that the voice-operated speaker is likely to be the father. Information characteristic of such a user is registered in advance in the registered user information 330 in association with the direction information.

The voice operation method in the voice operation device of the present embodiment is the flowchart of the first embodiment shown in FIG. 3 above and the second embodiment shown in FIG. 5 above, except that the auxiliary information is direction information. Both form flowcharts are applicable.

As described above, the voice operation device of the present embodiment further includes an array microphone that acquires the direction of the speaker of the voice operation as direction information. Then, the speaker identification unit identifies the speaker using the direction information. Even with such a configuration, it is possible to provide a voice operation device and a control method thereof that can further improve the accuracy of speaker identification. Instead of further providing the array microphone 42, the microphone 2 may be configured by a plurality of microphones to specify the direction of the speaker of the voice operation.

(Fourth Embodiment)
FIG. 7 is a block diagram schematically showing the configuration of the voice operation device according to the fourth embodiment. The voice operation device of this embodiment is different from the first embodiment in that it includes a distance measuring sensor 43 and a distance calculation unit 343. Other configurations are substantially the same as those in the first embodiment. Hereinafter, a configuration different from that of the first embodiment will be mainly described.

The distance measuring sensor 43 can acquire the distance to the speaker. As the distance measuring method, for example, a method of measuring a distance using a parallax or a well-known technique of measuring a distance using a reflected wave of light or a radio wave can be used. The distance calculation unit 343 calculates the distance from the voice control device to the speaker based on the information input from the distance measurement sensor 43. Here, the distance calculation unit 343 may be configured to be included in the distance measurement sensor 43. As described above, the voice operation device of the present embodiment is characterized by including the distance calculation unit 343 as the auxiliary information calculation unit 34 for calculating the auxiliary information.

For example, in the room shown in FIG. 2, the chair farther from the voice control device is the seat for the mother, and the chair closer to the voice control device is the seat for the father. In this case, the speaker identification unit 33 of the present embodiment determines that, for example, when the distance from the voice operation device to the speaker is long, the speaker of the voice operation is likely to be a mother. Alternatively, when the distance from the voice control device to the speaker is short, it is determined that the voice control speaker is likely to be the father. Information characteristic of such a user is registered in advance in the registered user information 330 in association with the distance information.

The voice operation method in the voice operation device of the present embodiment is the flowchart of the first embodiment shown in FIG. 3 above and the second embodiment shown in FIG. 5 above, except that the auxiliary information is distance information. Both form flowcharts are applicable.

As described above, the voice operation device of the present embodiment further includes a distance measuring sensor that acquires the distance to the speaker of the voice operation as distance information. Then, the speaker identification unit identifies the speaker using the distance information. Even with such a configuration, it is possible to provide a voice operation device and a control method thereof that can further improve the accuracy of speaker identification.

(Fifth Embodiment)
FIG. 8 is a block diagram schematically showing the configuration of the voice operation device according to the fifth embodiment. The voice operation device of the present embodiment is different from the first embodiment in that it includes a clock 44 and a time calculation unit 344. Other configurations are substantially the same as those in the first embodiment. Hereinafter, a configuration different from that of the first embodiment will be mainly described.

The clock 44 can acquire the current time. The time calculation unit 344 calculates the utterance time of the voice operation based on the information input from the clock 44. Here, the time calculation unit 344 may be configured to be included in the clock 44. As described above, the voice operation device of the present embodiment is characterized by including the time calculation unit 344 as the auxiliary information calculation unit 34 for calculating the auxiliary information.

For example, when the utterance time of the voice operation is daytime, the speaker identification unit 33 of the present embodiment determines that the speaker of the voice operation is likely to be a mother. Alternatively, when the utterance time of the voice operation is midnight, it is determined that the speaker of the voice operation is likely to be the father. Information characteristic of such a user is registered in advance in the registered user information 330 in association with the time information.

The voice operation method in the voice operation device of the present embodiment is the flowchart of the first embodiment shown in FIG. 3 above and the second embodiment shown in FIG. 5 above, except that the auxiliary information is time information. Both form flowcharts are applicable.

As described above, the voice operation device of the present embodiment further includes a clock that acquires the utterance time of the voice operation as time information. Then, the speaker identification unit identifies the speaker using the time information. Even with such a configuration, it is possible to provide a voice operation device and a control method thereof that can further improve the accuracy of speaker identification.

(Sixth Embodiment)
FIG. 9 is a block diagram schematically showing the configuration of the voice operation device according to the sixth embodiment. The voice operation device of the present embodiment has a configuration in which the configurations of the above-described embodiments are combined. The auxiliary information acquisition device 40 shown in FIG. 9 includes at least one of the GPS device 41, the array microphone 42, the distance measuring sensor 43, and the clock 44 of the above-described embodiment. Further, the auxiliary information calculation unit 34 shown in FIG. 9 is the auxiliary information among the voice operation recognition unit 32, the position calculation unit 341, the voice direction calculation unit 342, the distance calculation unit 343, and the time calculation unit 344 of the above-described embodiment. It includes at least a configuration capable of processing auxiliary information output by the acquisition device 40. Here, the auxiliary information calculation unit 34 may be configured to be included in the auxiliary information acquisition device 40.

The voice operation method in the voice operation device of the present embodiment is described in FIG. 3 and above, except that the auxiliary information is a combination of the voice operation information, the position information, the direction information, the distance information, and the time information of the above-described embodiment. The flowchart shown in FIG. 5 is applicable to both.

As described above, the voice operation device of the present embodiment includes a speaker identification unit that identifies the user as the speaker of the voice operation based on the voice information and the voice quality model of the user registered in advance. Then, the speaker identification unit identifies the speaker by using at least one of voice operation information, position information of the voice operation device, direction information of the speaker, distance information of the speaker, and time information as auxiliary information. .. According to such a configuration, it is possible to provide a voice operation device capable of further improving the accuracy of speaker identification and a control method thereof.

(Other embodiments)
It should be noted that the above-described embodiments are merely examples of embodiment of the present invention, and the technical scope of the present invention should not be construed in a limited manner by these. That is, the present invention can be implemented in various forms without departing from the technical idea or its main features.

For example, the user's schedule information may be registered in the registered user information 330 in advance, and the similarity may be corrected according to the schedule information. As an example, when the father is on a business trip in the schedule information, the similarity 1 / s between the voice information input from the voice input unit 21 and the voice quality model 331 of the father is set to -5 points.

Alternatively, the volume of the voice characteristic of the user may be registered in the registered user information 330 in advance, and the similarity may be corrected according to the volume of the voice of the voice operation. As an example, when the voice of the voice operation is loud, the similarity 1 / s between the voice information input from the voice input unit 21 and the voice quality model 331 of the father is +5 points.

Information such as the direction and distance in which the user is likely to exist may be registered in advance in the registered user information 330 in association with the auxiliary information, or the voice operating device may learn at any time. .. When the voice operation device learns, the correction value of the similarity is registered in advance in the registered user information 330 for each user, and the result of identifying the speaker is reflected in the correction value. Alternatively, the voice operation device constantly collects information such as keywords, positions, directions, distances, and times issued by the user, and updates information characteristic to the user of the registered user information 330 according to the information. You may do so. As a result, the speaker identification unit 33 learns the correlation between the information characteristic of the user registered in the registered user information 330 and the auxiliary information input via the auxiliary information calculation unit 34. Become. Therefore, the speaker identification unit 33 can improve the accuracy of speaker identification each time the operation process corresponding to the voice operation is performed.

Further, when a plurality of types of auxiliary information are used together as in the sixth embodiment, the auxiliary information used for speaker identification or a combination thereof may be selected according to the position and direction of the user. In this case as well, information characteristic of the user, such as the position and direction in which the user has a high probability of existence, is registered in the registered user information 330 in advance, and auxiliary information used for speaker identification is selected according to this information. Alternatively, the voice control device may learn at any time.

Further, in the above-described first to sixth embodiments, both the flowchart of the first embodiment shown in FIG. 3 and the flowchart of the second embodiment shown in FIG. 5 can be applied. That is, a method of registering a plurality of voice quality models 331 for each user and selecting the voice quality model 331 according to the auxiliary information, and registering information characteristic of the user in association with the auxiliary information to assist. Any method of correcting the similarity according to the information may be used.

Some or all of the above embodiments may also be described, but not limited to:

(Appendix 1)
A speaker identification unit that identifies the user as a speaker for voice operation based on voice information and a user's voice quality model registered in advance.
A voice operation recognition unit that recognizes the voice information and generates voice operation information,
It is a voice operation device equipped with
The speaker identification unit identifies the speaker by using at least one of the voice operation information, the position information of the voice operation device, the direction information of the speaker, the distance information of the speaker, and the time information as auxiliary information. Voice control device.

(Appendix 2)
A microphone that converts voice vibration into the voice information,
Registered user information in which the voice quality model of the user is registered in advance, and
The voice operation device according to Appendix 1.

(Appendix 3)
The registered user information has a plurality of the voice quality models for each user.
The voice operation device according to Appendix 2, wherein the speaker identification unit selects the voice quality model according to the auxiliary information.

(Appendix 4)
The voice operation device according to Appendix 3, wherein the speaker identification unit selects the voice quality model, which is a database for each user, according to the auxiliary information.

(Appendix 5)
The voice operation device according to any one of Supplementary note 2 to 4, wherein the speaker identification unit calculates the similarity between the voice information and the voice quality model, and identifies the speaker based on the similarity.

(Appendix 6)
The voice operation device according to Appendix 5, wherein the speaker identification unit identifies the user having the highest degree of similarity as a speaker for voice operation.

(Appendix 7)
The registered user information has information characteristic of the user in association with the auxiliary information.
The voice operation device according to Appendix 5 or 6, wherein the speaker identification unit corrects the similarity according to the auxiliary information.

(Appendix 8)
The registered user information has a correction value of the similarity for each user.
The voice operation device according to any one of Appendix 5 to 7, wherein the speaker identification unit reflects the result of identifying the speaker in the correction value and learns the correlation between the auxiliary information and the correction value. ..

(Appendix 9)
The registered user information includes keywords characteristic of the user, and includes keywords that are characteristic of the user.
The voice operation device according to any one of Appendix 5 to 8, wherein the speaker identification unit corrects the similarity according to the keyword included in the voice operation information.

(Appendix 10)
The microphone is an array microphone that acquires the direction of a speaker for voice operation as the direction information.
The voice operation device according to any one of Supplementary note 2 to 9, wherein the speaker identification unit identifies a speaker by using the direction information.

(Appendix 11)
Further equipped with a distance measuring sensor that acquires the distance to the speaker of the voice operation as the distance information,
The voice operation device according to any one of Supplementary note 2 to 10, wherein the speaker identification unit identifies a speaker by using the distance information.

(Appendix 12)
A clock for acquiring the utterance time of the voice operation as the time information is further provided.
The voice operation device according to any one of Supplementary note 2 to 11, wherein the speaker identification unit identifies a speaker by using the time information.

(Appendix 13)
A GPS device that acquires the position of the voice control device as the position information is further provided.
The voice operation device according to any one of Appendix 2 to 12, wherein the speaker identification unit selects the voice quality model according to the position information.

(Appendix 14)
The voice operation device according to any one of Supplementary note 2 to 13, wherein the registered user information further includes the schedule information of the user, and the speaker identification unit further uses the schedule information to identify the speaker.

(Appendix 15)
The registered user information has information according to the preference of the user for each user.
A control calculation unit that executes operation processing corresponding to the voice operation information according to the preference of the user, and
A speaker that notifies the execution result by voice and
The voice operating device according to any one of Supplementary note 2 to 14.

(Appendix 16)
A speaker identification unit that identifies the user as a speaker for voice operation based on voice information and a user's voice quality model registered in advance.
A voice operation recognition unit that recognizes the voice information and generates voice operation information,
It is a control method of a voice operation device equipped with
The speaker identification unit identifies the speaker by using at least one of the voice operation information, the position information of the voice operation device, the direction information of the speaker, the distance information of the speaker, and the time information as auxiliary information. A method of controlling a voice control device having steps.

Although the present invention has been described above with reference to the embodiments, the present invention is not limited to the above embodiments. Various changes that can be understood by those skilled in the art can be made to the structure and details of the present invention within the scope of the present invention.

This application claims priority on the basis of Japanese application Japanese Patent Application No. 2017-144336 filed on July 26, 2017, and incorporates all of its disclosures herein.

1: Speaker 2: Microphone 3: Control calculation unit 11: Voice output unit 21: Voice input unit 31: Voice operation response unit 32: Voice operation recognition unit 33: Speaker identification unit 34: Auxiliary information calculation unit 35: Wireless communication unit 40: Auxiliary information acquisition device 41: GPS device 42: Array microphone 43: Distance measurement sensor 44: Clock 61: User 62: User 330: Registered user information 331: Voice quality model 341: Position calculation unit 342: Voice direction calculation Unit 343: Distance calculation unit 344: Time calculation unit

Claims (11)

A speaker identification unit that identifies the user as a speaker for voice operation based on voice information and a user's voice quality model registered in advance.
A voice operation recognition unit that recognizes the voice information and generates voice operation information,
GPS device and
A microphone that converts voice vibration into the voice information,
Registered user information in which the voice quality model of the user is registered in advance, and
It is a voice operation device equipped with
The registered user information has a correction value of similarity between the voice information and the voice quality model for each user.
The speaker identification unit assists by using at least one of the voice operation information, the position information of the voice operation device acquired by the GPS device, the direction information of the speaker, the distance information of the speaker, and the time information. A voice operation device that identifies a speaker using information and the similarity , reflects the identification result in the correction value, and learns the correlation between the auxiliary information and the correction value. The auxiliary information is at least one of the position information of the voice operating device, the direction information of the speaker, the distance information of the speaker, and the time information.
The registered user information has a plurality of the voice quality models for each user.
The speaker identification unit selects the voice quality model from the plurality of voice quality models according to the auxiliary information, and makes the user perform the voice operation based on the voice information and the selected voice quality model. The voice operating device according to claim 1, which identifies the speaker. The voice operation device according to claim 2 , wherein the speaker identification unit selects the voice quality model, which is a database for each user, according to the auxiliary information. The voice operation device according to any one of claims 1 to 3, wherein the speaker identification unit identifies the user having the highest degree of similarity as a speaker for voice operation. The registered user information has information characteristic of the user in association with the auxiliary information.
The voice operation device according to any one of claims 1 to 4, wherein the speaker identification unit corrects the similarity according to the auxiliary information. The registered user information includes keywords characteristic of the user, and includes keywords that are characteristic of the user.
The voice operation device according to any one of claims 1 to 5 , wherein the speaker identification unit corrects the similarity according to the keyword included in the voice operation information. The microphone is an array microphone that acquires the direction of a speaker for voice operation as the direction information.
The voice operation device according to any one of claims 1 to 6 , wherein the speaker identification unit identifies a speaker by using the direction information. Further equipped with a distance measuring sensor that acquires the distance to the speaker of the voice operation as the distance information,
The voice operation device according to any one of claims 1 to 7 , wherein the speaker identification unit identifies a speaker by using the distance information. A clock for acquiring the utterance time of the voice operation as the time information is further provided.
The voice operation device according to any one of claims 1 to 8 , wherein the speaker identification unit identifies a speaker by using the time information. The voice operation device according to any one of claims 1 to 9 , wherein the registered user information further includes a schedule information of the user, and the speaker identification unit further uses the schedule information to identify a speaker. The registered user information has information according to the preference of the user for each user.
A control calculation unit that executes operation processing corresponding to the voice operation information according to the preference of the user, and
A speaker that notifies the execution result by voice and
Further voice operation device according to any one of claims 1 1 0 having.

Images